Re: Issue with acpi0 on Intel NUC11TNHi3

On 1/16/23 03:01, Bradley Latus wrote:
> Hello Stuart,
>
> I noticed that someone else had a similar issue on the openbsd-bugs list..
> https://marc.info/?l=openbsd-bugs&m=166497715729842&w=2
>
> I was able to apply a patch I found from another user (Joe Miller)
> which masks out
> GPE_L6F messages and the problem was resolved.
> https://gist.github.com/joemiller/9f5698c5634d4a93d101985dc5238365
> https://news.ycombinator.com/item?id=33279037
>
> After applying his patch (removing the additional printing parts)
> My system was restored to what should be expected.
>

I can confirm the issue, cpu0 mainly busy with something ACPI related,
cpu1 spinning a lot, the whole system being slow, barely usable.

Applying the patch to mask/ignore the offending interrupt appears to
solve the problem for me.

regards,
Remco


Index: acpi.c
===================================================================
RCS file: /cvs/src/sys/dev/acpi/acpi.c,v
retrieving revision 1.418
diff -u -p -u -p -r1.418 acpi.c
--- acpi.c 13 Sep 2022 17:14:54 -0000 1.418
+++ acpi.c 18 Jan 2023 17:32:34 -0000
@@ -2300,6 +2300,12 @@ acpi_gpe(struct acpi_softc *sc, int gpe,
struct aml_node *node = arg;
uint8_t mask, en;

+ /* mask/ignore the GPE_L6F interrupt */
+ if (gpe == 0x6f) {
+ if ( (sc->gpe_table[gpe].flags & GPE_LEVEL) )
+ return (0);
+ }
+
dnprintf(10, "handling GPE %.2x\n", gpe);
aml_evalnode(sc, node, 0, NULL, NULL);